Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
NCTM or The Math Forum.



Re: Arithmetic compression
Posted:
Jul 22, 2014 11:51 AM


jonas.thornvall@gmail.com writes:
> Den tisdagen den 22:e juli 2014 kl. 16:20:37 UTC+2 skrev Ben Bacarisse: >> jonas.thornvall@gmail.com writes: >> >> > Den tisdagen den 22:e juli 2014 kl. 14:06:40 UTC+2 skrev Ben Bacarisse: >> >> jonas.thornvall@gmail.com writes: >> >> <snip> >> >> > I actually have a arithmetic circuit switch in mind that will save >> >> > alot of digit compressing binary raw data rather then data types, >> >> > roughly a third will be compressed. >> >> >> >> There are two keys things about any compression scheme: the set of >> >> inputs that get compressed (and by how much), and the set of inputs that >> >> expand (and by how much). The most useful schemes have frequently >> >> occurring data in the first and rare data in the second. >> >> >> > It will a binary sequense of any length by a third. >> >> What lengths will the sequences "", "0", "1", "00", "01", "10" and "11 >> become? > > I do not think anyone but a fool is interested in compressing binary > digit pairs. Are you a fool?
Yes, I am, but my point still stands. All you need to do is rephrase "binary sequence of any length" so that it becomes true for your proposed algorithm. (I can't help with that because I don't know the algorithm).
 Ben.



